Launch Protocol: How Law Firms Are Adopting AI in Real Time

Summary Jared Correia sits down with Sean McTigue, a partner at Bartko Pavia LLP and one of the more technically fluent attorneys in practice today. Sean unpacks how his firm navigated the leap from legal-specific AI tools to a direct enterprise deployment of Anthropic's models, and why he thinks that distinction matters a lot more than most firms realize. The conversation covers practical ground: how to use Westlaw's Quickcheck as a verification loop, why lawyers overestimate what AI will do for them on the first try, and how to find the early adopters inside a firm and turn their discoveries into firm-wide workflows. Sean also looks ahead at what AI means for the billable hour model and why the legal profession can't afford to stay in the way. About the Guest Sean McTigue is a partner at Bartko Pavia LLP in San Francisco, where he handles complex litigation with a particular focus on integrating AI into the practice of law. He has been following the development of large language models closely since GPT-4's launch and has led the firm's rollout of Anthropic for Enterprise. Sean studied philosophy at the University of Utah and earned his law degree at Berkeley Law. Key Takeaways Hallucination risk in AI outputs is a solved problem, using tools like Westlaw's Quickcheck as a verification flywheel alongside AI drafting, not a reason to avoid AI entirely. Legal-specific tools rarely add value beyond a general foundation model; the wrapper around the model matters less than most vendors claim. Direct enterprise deployment of a foundation model lets firms ride the frontier rather than being stuck on whatever model a SaaS vendor last tested. The billable hour model is under pressure, and firms that build internal AI capital now are better positioned to shift toward fixed-fee and alternative-fee arrangements. Adoption inside a firm starts with finding the heavy users, learning what they figured out, and distributing those workflows to everyone else. Links and Resources Red Cave Law Firm Consulting Bartko Pavia LLP Westlaw CoCounsel Westlaw Quickcheck - available inside your Westlaw subscription Anthropic for Enterprise Keywords AI adoption in law firms, legal AI tools, Westlaw Quickcheck, AI hallucinations legal, foundation models for lawyers, Anthropic for Enterprise, billable hours AI, legal tech vendor evaluation, Sean McTigue, Bartko Pavia, Jared Correia, Adventures in Legal Tech, CoCounsel Westlaw, AI verification legal, small firm AI, legal workflow automation, enterprise AI deployment, AI research tools lawyers, prompt engineering legal, alternative fee arrangements AI Episode Highlights [00:02:01 - 00:04:54] Sean introduces Westlaw Quickcheck as the underused verification tool that turns hallucination risk into a manageable step in the workflow. [00:05:00 - 00:07:57] Sean explains why lawyers who try AI once, find it imperfect, and dismiss it are missing the workflow question entirely. [00:08:13 - 00:09:28] The hammer-and-nail analogy: being handed a tool and told to use it without any guidance on what the full project actually looks like. [00:19:14 - 00:23:27] Sean describes the frustration of vetting legal AI vendors who can't tell you what model they're running, including an e-discovery platform using Haiku 3 on million-document reviews. [00:24:54 - 00:28:05] The case for direct foundation model deployment over legal-specific SaaS wrappers, and what you can do with a generalist model that a niche tool will never offer. [00:36:44 - 00:40:43] The future of legal billing: from the billable hour back toward fixed-fee engagements, and why firms that build AI capital now are better positioned. [00:41:50 - 00:47:23] Sean's starter recommendation: Westlaw citing references downloaded in bulk and fed to an LLM, plus why Google AI Overview is already AI whether lawyers know it or not.